Pierfrancesco Lijoi

Architecting a cloud-native defect prediction SaaS on AWS microservices using Infrastructure as Code (Terraform) to codify and manage EC2 infrastructure in a scalable and reproducible way. Building production-grade Java/Python services integrating LLM-based semantic analysis and ML models for automated code risk detection. Owning end-to-end ML pipelines: data collection, cleaning, exploratory analysis, training, evaluation, and feature engineering for le-level defect prediction. Automating testing and deployment via CI/CD (GitHub Actions), enabling continuous risk assessment on every pull request. Packaging the GitHub Actions tool using Docker-based workflows to ensure consistent execution, portability, and ready-to-use developer integration. Shipped a production-ready GitHub Actions integration for automated PR risk analysis, reaching 80% precision and 40% recall on le-level defect prediction.

Delivered MongoDB/NoSQL engineering training to enterprise development teams, accelerating adoption of modern database practices. Designed hands-on labs and e-learning modules simulating real-world scalability and performance challenges. Enabled teams to migrate toward cloud-ready and scalable architectures, improving system maintainability and reliability.

Led requirements analysis and system design, translating business needs into scalable and maintainable software architectures. Developed cross-platform applications using C#, Python, .NET MAUI, and Azure Functions, applying the MVVM pattern to enforce separation of concerns, testability, and clean architecture. Improved system reliability and performance through automated testing, systematic debugging, and refactoring, reducing production issues and maintenance overhead.